Noya.ai Research Report: The Revolution of Prediction Market Agents under the AI×DeFi Narrative

Author: 0xjacobzhao

In previous Crypto AI series research reports, we have consistently emphasized the following view: the most practically valuable scenarios in the current crypto space are mainly concentrated on Stablecoin Payments and DeFi, with Agent being the key interface for AI industry users. Therefore, in the trend of Crypto and AI integration, the two most valuable paths are: a short-term AgentFi based on existing mature DeFi protocols (such as lending, liquidity mining, and other basic strategies, as well as advanced strategies like Swap, Pendle PT, and funding rate arbitrage), and a medium- to long-term focus on stablecoin settlement, relying on protocols like ACP/AP2/x402/ERC-8004 to develop Agent Payment.

Prediction markets have become an industry trend that cannot be ignored by 2025, with annual total trading volume skyrocketing from about $9 billion in 2024 to over $40 billion in 2025, achieving over 400% year-over-year growth. This significant growth is driven by multiple factors: macro political events (such as the 2024 US election) creating demand for uncertainty, infrastructure and trading model maturity, and regulatory breakthroughs (Kalshi’s lawsuit victory and Polymarket’s return to the US). The Prediction Market Agent(Prediction Market Agent) is expected to take shape in early 2026 and may become an emerging product form in the agent field within the next year.

1. Prediction Markets: From Betting Tools to the “Global Truth Layer”

Prediction markets are financial mechanisms that facilitate trading based on future event outcomes, with contract prices essentially reflecting the market’s collective judgment of the probability of events. Their effectiveness stems from the combination of crowd wisdom and economic incentives: in environments where anonymous, real-money betting occurs, dispersed information is rapidly integrated into price signals weighted by capital willingness, significantly reducing noise and false judgments.

By the end of 2025, prediction markets have basically formed a duopoly dominated by Polymarket and Kalshi. According to Forbes, the total trading volume in 2025 reached approximately $44 billion, with Polymarket contributing about $21.5 billion and Kalshi about $17.1 billion. Kalshi’s rapid expansion is attributed to its legal victory in previous election contracts, its early compliance advantage in the US sports prediction market, and relatively clear regulatory expectations. Currently, their development paths are clearly diverging:

• Polymarket adopts a hybrid CLOB architecture with “off-chain matching, on-chain settlement” and a decentralized settlement mechanism, building a global, non-custodial, highly liquid market. After returning to US compliance, it operates a “onshore + offshore” dual-track structure.
• Kalshi integrates into traditional financial systems, connecting via API to mainstream retail brokers, attracting Wall Street market makers to deeply participate in macro and data-based contracts. Its products are constrained by traditional regulatory processes, with long-tail demand and sudden events lagging behind.

Besides Polymarket and Kalshi, other competitive participants in prediction markets mainly develop along two paths:

• Compliance distribution path: embedding event contracts into existing broker or large platform account systems, leveraging channel coverage, clearing capacity, and institutional trust (e.g., ForecastTrader via Interactive Brokers and ForecastEx, FanDuel with CME).
• On-chain performance and capital efficiency path: exemplified by Solana’s perpetual contract DEX Drift, which added a prediction market module B.E.T (prediction markets) to its existing product line.

The combination of traditional financial compliance entry points and native crypto performance advantages forms a diverse competitive landscape in prediction markets.

Prediction markets superficially resemble gambling, but fundamentally are a zero-sum game. The core difference lies not in form but in whether they generate positive externalities: aggregating dispersed information through real-money trading to publicly price future events, forming valuable signals. Despite limitations like entertainment participation, the trend is shifting from gaming toward the “global truth layer”—with institutions like CME and Bloomberg providing access, event probabilities have become decision-making metadata directly callable by financial and corporate systems, offering more timely and quantifiable market-based truths.

2. Prediction Intelligence Agents: Architecture, Business Models, and Strategy Analysis

Currently, Prediction Market Agents(Prediction Market Agent) are entering early practice stages. Their value is not in “more accurate AI predictions,” but in amplifying information processing and execution efficiency within prediction markets. Prediction markets are essentially information aggregation mechanisms, with prices reflecting collective judgments of event probabilities; market inefficiencies stem from information asymmetry, liquidity, and attention constraints. The rational positioning of prediction market agents is Executable Probabilistic Portfolio Management: converting news, rule texts, and on-chain data into verifiable pricing deviations, enabling faster, more disciplined, and low-cost strategy execution, capturing structural opportunities through cross-platform arbitrage and portfolio risk management.

An ideal prediction market agent can be abstracted into a four-layer architecture:

• Information Layer: gathers news, social data, on-chain and official data;
• Analysis Layer: uses LLM and ML to identify mispricings and compute edges;
• Strategy Layer: converts edges into positions via Kelly formulas, batch building, and risk controls;
• Execution Layer: completes multi-market order placement, slippage and gas optimization, and arbitrage execution, forming an efficient automated closed loop.

The ideal business model design for prediction market agents has exploration space at different levels:

• Infrastructure Layer: provides multi-source real-time data aggregation, Smart Money address databases, unified prediction market execution engines, and backtesting tools, charging B2B/B2D, ensuring stable income unrelated to prediction accuracy;
• Strategy Layer: deposits modular strategy components and community-contributed strategies via open-source or Token-Gated mechanisms, forming a composable strategy ecosystem and capturing value;
• Agent Layer: directly runs real trading through entrusted vaults, with transparent on-chain records and 20–30% performance fees (plus small management fees), demonstrating capability.

A predictive market intelligent agent is closer to an “AI-driven probabilistic asset management product,” executing long-term disciplined strategies and mispricing arbitrage across markets, rather than relying on single prediction accuracy for profit. The core logic of “Infrastructure monetization + Ecosystem expansion + Performance participation” is that, even as alpha converges with market maturity, underlying capabilities like execution, risk control, and settlement retain long-term value, reducing dependence on the assumption of “AI continuously beating the market.”

Strategy Analysis of Prediction Market Agents:

In theory, agents have advantages in high-speed, 24/7, de-emotionalized execution, but in prediction markets, they often struggle to generate sustained alpha. Their effective application is mainly limited to specific structures such as automated market making, cross-platform mispricing capture, and long-tail event information aggregation—opportunities that are scarce and constrained by liquidity and capital.

1. Market Selection: Not all prediction markets are tradable; participation value depends on clarity of settlement, liquidity quality, informational advantage, timing structure, and manipulation risk. Prioritize early-stage new markets, long-tail events with few professional players, and short-lived pricing windows caused by time zone differences; avoid high-profile political events, subjective settlement markets, and very low liquidity assets.
2. Order Placement Strategy: Use strict systematic position management. Enter only when your probability judgment significantly exceeds market implied probability, based on Fractional Kelly (usually 1/10–1/4 Kelly). Limit individual event risk exposure to under 15%, aiming for risk-controlled, drawdown-tolerant, advantage-compounding steady growth.
3. Arbitrage Strategies: Main arbitrage types in prediction markets include: Cross-platform price differences (beware of settlement discrepancies), Dutch Book arbitrage (high certainty but strict liquidity requirements), Settlement arbitrage (dependent on execution speed), and Related asset hedging (limited by structural mismatches). The key is not just identifying price differences but strictly aligning contract definitions and settlement standards to avoid pseudo-arbitrage caused by rule nuances.
4. Smart Money Copying: On-chain “smart money” signals are lagging, prone to inducement risks and sample bias, and are not suitable as primary strategies. They are better used as confidence adjustment factors to assist core judgments based on information and pricing deviations.

3. Noya.ai: From Intelligence to Action in the Agent Network

As an early exploration of prediction market agents, NOYA’s core idea is “Intelligence That Acts”. In on-chain markets, mere analysis and insights are insufficient to create value—although dashboards, data analysis, and research tools help users understand “what might happen,” there are still many manual operations, cross-chain frictions, and execution risks between insight and action. NOYA addresses this pain point by building a complete chain: “Research → Judgment → Execution → Continuous Monitoring,” compressed into a unified system that enables intelligence to directly translate into on-chain actions.

NOYA achieves this by integrating three core layers:

• Intelligence Layer (Intelligence): aggregates market data, token analysis, and prediction market signals.
• Abstraction Layer (Abstraction): hides complex cross-chain routing, allowing users to express Intent.
• Execution Layer (Execution): AI Agents execute operations across chains and protocols based on user authorization.

In product form, NOYA supports passive income users, active traders, and prediction market participants, with designs like Omnichain Execution, AI Agents & Intents, Vault Abstraction, modularizing and automating multi-chain liquidity management, complex strategy execution, and risk control.

The overall system forms a continuous closed loop: Intelligence → Intent → Execution → Monitoring, ensuring users always retain control over assets while achieving efficient, verifiable, low-friction transition from insight to execution.

4. Noya.ai Product System and Evolution Path

Core Foundation: Noya Omnichain Vaults

Omnivaults are NOYA’s capital deployment layer, offering cross-chain, risk-controlled automated yield strategies. Users deposit assets with simple operations, entrusting the system to continuously operate across multiple chains and protocols without manual rebalancing or monitoring. The core goal is to achieve stable risk-adjusted returns rather than short-term speculation.

Omnivaults cover standard yield and looping strategies, clearly categorized by asset and risk levels, supporting optional incentive binding mechanisms. The system automatically handles cross-chain routing and optimization, with the potential to incorporate ZKML for verifiable strategy proofs, enhancing transparency and trustworthiness. The design emphasizes modularity and composability, supporting future integration of more asset types and advanced strategies.

NOYA Vault (金库) Technical Architecture: Each vault is registered and managed via Registry, with AccountingManager handling user shares (ERC-20) and net asset value pricing; underlying modules Connectors interface with protocols like Aave, Uniswap, and others, calculating cross-protocol TVL, relying on Value Oracle (Chainlink + Uniswap v3 TWAP) for pricing and valuation; trading and cross-chain operations are executed via Swap Handler (LiFi); final strategy execution is triggered by Keeper multisig, forming a auditable, composable execution loop.

Future Alpha: Prediction Market Intelligent Agents (Prediction Market Agent)

NOYA’s most imaginative module: Intelligence layer continuously tracks on-chain capital behavior and off-chain narrative shifts, identifying news shocks, sentiment swings, and odds mismatches. When probability deviations are found in prediction markets like Polymarket, the execution layer’s AI Agent can, with user authorization, mobilize vault funds for arbitrage and rebalancing. Additionally, Token Intelligence and Prediction Market Copilot provide structured token and prediction market analysis, directly translating external information into actionable trading decisions.

Prediction Market Intelligent Decision-Making Assistant (Prediction Market Intelligence Copilot))

NOYA aims to upgrade prediction markets from single-event bets to systematically managed probabilistic assets. Its core modules integrate implied probabilities, liquidity structures, historical settlements, and on-chain smart money signals, using EV and scenario analysis to identify pricing deviations. It also tracks high-win-rate wallet positions to distinguish information trading from market noise. Based on this, Copilot supports cross-market, cross-event correlation analysis, transmitting real-time signals to AI Agents for automated opening and adjusting positions, enabling portfolio management and dynamic optimization of prediction markets.

Core strategy mechanisms include:

• Multi-source Edge information capture: combining real-time odds from Polymarket, polls, private and external data streams, cross-validating event implied probabilities, systematically uncovering undervalued information advantages.
• Cross-market and cross-event arbitrage: exploiting pricing differences across markets, contract structures, or similar events to build probability and structural arbitrage strategies, controlling directional risk to capture odds convergence gains.
• Odds-driven dynamic position management: when odds shift significantly due to information, capital, or sentiment changes, AI Agents automatically adjust positions, maintaining continuous optimization rather than one-time bets.

NOYA Token Intelligence Reports (NOYA Intelligence Token Reports)

NOYA’s institutional research and decision core aims to automate professional crypto research workflows and produce decision-level signals for real asset allocation. The module presents standardized reports with clear investment stances, comprehensive scores, core logic, key catalysts, and risk warnings, continuously updated with real-time market and on-chain data. Unlike traditional research tools, NOYA’s intelligence is not static; it can be called, compared, and queried via AI Agents in natural language, directly feeding into execution layers to drive cross-chain trading, capital allocation, and portfolio management, forming an integrated “research—decision—execution” closed loop, making Intelligence an active signal source in automated capital operations.

NOYA AI Agent (Voice and Natural Language Driven)

NOYA AI Agent is the platform’s execution layer, primarily responsible for transforming user intent and market intelligence into authorized on-chain actions. Users can express goals via text or voice, and the Agent plans and executes cross-chain, cross-protocol operations, compressing research and execution into a seamless process. It is a key product for lowering operational barriers in DeFi and prediction markets.

Users do not need to understand underlying chains, protocols, or transaction paths—simply expressing goals in natural language or speech triggers the AI Agent to automatically plan and execute multi-step on-chain operations, realizing “intent equals execution.” Under full user signature and non-custodial conditions, the Agent operates in a closed loop of “understanding intent → planning action → user confirmation → on-chain execution → result monitoring,” not replacing decision-making but ensuring efficient implementation, significantly reducing friction and complexity in advanced financial operations.

Trust moat: ZKML Verifiable Execution

Verifiable execution aims to establish a fully auditable, trust-minimized process from strategy to decision to execution. NOYA introduces ZKML as a key mechanism to reduce trust assumptions: strategies are computed off-chain with verifiable proofs, and only upon on-chain verification are corresponding fund operations triggered. This mechanism can provide strategy credibility without revealing model details and support verifiable backtesting and other capabilities. The related modules are still marked as “in development” in public documents, with engineering details to be disclosed later.

Next 6 months product roadmap

• Enhance prediction market advanced order capabilities to support agent-based trading.
• Expand to more prediction platforms beyond Polymarket to increase event coverage and liquidity.
• Improve multi-source Edge information collection, cross-validating with odds data to systematically capture undervalued probabilities.
• Deliver clearer token signals and high-level reports to directly drive trading signals and on-chain analysis.
• Launch more sophisticated on-chain DeFi strategy combinations to improve capital efficiency, yield, and scalability.

5. Noya.ai Ecosystem Growth and Incentive System

Currently, Omnichain Vaults are in early ecosystem development, with cross-chain execution and multi-strategy frameworks validated.

• Strategies and coverage: the platform has integrated major DeFi protocols like Aave and Morpho, supporting cross-chain deployment of stablecoins, ETH, and derivatives, with layered risk strategies (e.g., basic yield vs. Loop).
• Development stage: with limited TVL, the focus is on function validation (MVP) and risk control framework refinement, with a highly composable architecture that supports future integration of more assets and advanced agent scheduling.

Incentive system: Kaito linkage and Space Race dual-wheel drive

NOYA has built a “real contribution” anchored growth flywheel, deeply linking content narratives and liquidity.

1. Ecosystem cooperation (Kaito Yaps): NOYA’s “AI × DeFi × Agent” narrative is showcased on Kaito Leaderboards, with a 5% unvested incentive pool and an additional 1% for Kaito ecosystem. This mechanism tightly couples content creation (Yaps) with vault deposits and bond locking, turning weekly contributions into Stars that determine levels and multipliers, reinforcing narrative consensus and long-term capital stickiness.
2. Growth engine (Space Race): Space Race forms NOYA’s core growth flywheel, using Stars as long-term equity tokens, replacing traditional “fund size priority” airdrops. It combines bond locking bonuses, 10% bidirectional referral incentives, and content dissemination into weekly Points, selecting highly engaged, consensus-strong long-term users, continuously optimizing community structure and token distribution.
3. Community building (Ambassador): NOYA employs an invitation-based ambassador program, offering community participation rights and performance rebates (up to 10%) based on actual contributions.

Noya.ai has accumulated over 3,000 on-chain users, with X platform followers exceeding 41,000, ranking in the top five of Kaito Mindshare. This indicates NOYA has secured a favorable attention ecosystem in prediction markets and agent tracks.

Additionally, Noya.ai’s core contracts have undergone audits by Code4rena and Hacken, and are integrated with Hacken Extractor.

6. Token Economic Model and Governance

NOYA adopts a single-token ecosystem model, with $NOYA as the sole value bearer and governance token.

It employs a Buyback & Burn value capture mechanism, where value generated from AI Agents, Omnivaults, and Prediction Markets is captured via staking, governance, access rights, and buyback & burn mechanisms, creating a value loop of usage → fees → buyback & burn. This converts platform usage into long-term token value.

The project follows a Fair Launch principle, without angel or VC investments, instead distributing via public community rounds (Launch-Raise) at a low valuation ($10M FDV), Space Race, and airdrops, deliberately reserving asymmetric upside for the community, favoring active users and long-term participants; team incentives mainly come from long-term locked tokens.

Token Distribution (Distribution)

• Total supply: 1,000,000,000 NOYA
• Initial circulating supply (Low Float): approximately 10%
• Valuation & Fundraising (The Raise): Funding amount: $1 million; Valuation (FDV): $10 million

( 7. Prediction Market Agent Market Competition Analysis

Currently, the Prediction Market Agent track remains in early stages, with limited projects. Notable examples include Olas (Pearl Prediction Agents), Warden (BetFlix), and Noya.ai.

From product forms and user participation modes, these represent three paths in the prediction market agent space:

1. Olas (Pearl Prediction Agents): Productized and operationalized agents, encapsulating prediction market trading into “run an automated prediction agent”: users fund and run, and the system automatically handles information gathering, probability judgment, betting, and settlement. Additional installation requirements limit user-friendliness.

2. Warden (BetFlix): Interactive, consumer-grade betting platforms, attracting users through low-threshold, entertainment-focused experiences, emphasizing engagement and distribution. Uses gamification and content-driven frontends to lower participation barriers, highlighting prediction markets’ entertainment aspect. Its competitive advantage lies in user growth and distribution efficiency rather than deep strategy or execution.

3. NOYA.ai: Focused on “fund custody + strategy execution,” abstracting prediction markets and DeFi execution into asset management products via Vault, providing low-operation, low-mental-load participation. Future integration with Prediction Market Intelligence and Agent execution modules could form an “research—execution—monitoring” workflow. ![])https://img-cdn.gateio.im/webp-social/moments-8bb99f373c04010362ba48b8ae063152.webp###

Compared to Giza, Almanak, and other projects with clear product delivery, NOYA’s DeFi Agents are still relatively early. Its differentiation lies in positioning and entry level: it enters the same execution and asset management narrative with an approximate (FDV) valuation, offering significant valuation discount and growth potential.

• NOYA: Focused on Omnichain Vaults as core asset management encapsulation, with current delivery emphasizing cross-chain execution and risk control infrastructure. The upper-layer Agent execution, prediction market capabilities, and ZKML mechanisms are still under development and verification, not yet a mature DeFAI platform.

• Giza: Capable of directly running asset strategies (ARMA, Pulse), with the highest completion among AgentFi projects.

• Almanak: Positioned as AI Quant for DeFi, providing strategies and risk signals via models and quant frameworks, mainly targeting professional funds and strategy management, emphasizing methodology and reproducibility.

• Theoriq: Focused on multi-agent collaboration (Agent Swarms), emphasizing scalable agent cooperation and long-term infrastructure, more oriented toward foundational capability building.

• Infinit: A more execution-layer oriented Agentic DeFi terminal, orchestrating “intent → multi-step on-chain operations,” significantly lowering complex DeFi operation barriers, with user-perceived value being more direct.

$10M 8. Summary: Business Logic, Engineering, and Potential Risks

Business Logic:

NOYA is a relatively rare multi-narrative overlay combining AI Agent × Prediction Market × ZKML, further integrated with Intent-driven execution. Its asset pricing is initiated at about **###FDV$10M **, significantly below the common $75M–$100M valuation range of similar AI / DeFAI / Prediction projects, creating a structural valuation gap.

Design-wise, NOYA attempts to unify strategy execution (Vault/Agent) and information advantage (Prediction Market Intelligence) within the same execution framework, with protocol revenue flows (fees → buyback & burn) establishing a value capture loop. Although still early, the multi-narrative overlay and low valuation entry point give it a more asymmetric risk–reward profile akin to high-odds, non-symmetric betting assets.

Engineering implementation: In verifiable delivery, NOYA’s core functional deployment currently includes Omnichain Vaults, enabling cross-chain asset management, yield strategies, and delayed settlement. Its vision of Prediction Market Intelligence (Copilot), NOYA AI Agent, and ZKML-driven verifiable execution remains under development, not yet forming a complete on-chain closed loop. It is not yet a mature DeFAI platform.

Potential Risks and Focus Points:

1. Delivery Uncertainty: The leap from “basic Vault” to “full-featured Agent” involves significant technical scope; watch for delays or underperformance of ZKML implementation.
2. Systemic Risks: Contract security, cross-chain bridge failures, and prediction market-specific oracle disputes (e.g., rule ambiguities) could cause fund losses if any single point fails.

Disclaimer: This document was assisted by AI tools including ChatGPT-5.2, Gemini 3, and Claude Opus 4.5 during creation. The author has verified and strives for accuracy but may have omissions. Crypto markets often show divergence between fundamentals and secondary market prices. This content is for informational and research purposes only, not investment advice or token trading recommendations.